bioRxiv preprint doi: https://doi.org/10.1101/2020.08.22.262964; this version posted August 24, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 1 1 Plant traits that influence flower visits by birds in a montane forest 2 3 Oscar Gonzalez 4 Grupo Aves del Peru. Gomez del Carpio 135 Lima, Peru 5 and 6 Department of Natural Sciences, Emmanuel College, Franklin Springs 30662 GA, USA. 7 [email protected] 8 9 Abstract 10 In a bird-flowering plant network, birds select plants that present traits attractive 11 to them. I studied plant characteristics that might predict flower visitation rate by the 12 most common bird visitors in a bird-flowering plant network located in an elfin forest of 13 the Andes. The nectarivorous birds which had the highest number of interactions with 14 flowering plants in this network were the Coppery Metaltail (Metallura theresiae), the 15 Great Sapphirewing (Pterophanes cyanopterus), and the Moustached Flowerpiercer 16 (Diglossa mystacalis). I analyzed different flower traits (flower aggregation, nectar 17 volume, nectar energy, color, orientation, and dimensions of the corolla) of the common 18 plants that these birds visited with a principal component analysis. The plants most 19 visited by birds were Brachyotum lutescens and Tristerix longebracteatus. While nectar 20 traits of the plants seemed to be the best predictor for bird visitation, there was no 21 statistical association between visitation and plant traits, except for Metallura theresiae 22 in the dry season. I discuss the possible causes of resource partitioning for these 23 nectarivorous birds. 24 Keywords: Nectar, elfin forest, flower traits, nectarivorous birds 25 26 1 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.22.262964; this version posted August 24, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 2 27 Introduction 28 Birds that feed on nectar make decisions on multiple scales to select plants and 29 flowers; these scales could be at habitat, flowering patch, individual plant, or flower level 30 (Sutherland and Gass 1995; Ortiz-Pulido and Vargas-Licona 2008). The visitation of 31 each bird species may be different for the same resource (Feinsinger 1976; Davis et al. 32 2015). Different plant traits can attract flower visitors, such as the color of the corolla 33 (Wilson et al. 2006), the aggregation of flowers of the plants (Fonturbel et al. 2015), the 34 morphological matching of the feeding apparatus with the flower (Cotton 2007), flower 35 orientation (Aizen 2003), or nectar properties. 36 Nectar is a primary resource for flower visitors and is a crucial determinant in 37 interactions between animals and plants (Wiens 1989, Rathcke 1992, Cotton 2007, 38 Janecek et al. 2012; Justino et al. 2012). The energy resource of nectar is determined 39 by volume present and sugar concentration; animals tend to preferentially visit flowers 40 with the most reward (Fleming et al. 2004). It is likely that nectarivorous birds - such as 41 hummingbirds or flowerpiercers - have specific preferences for some plants depending 42 on the nectar volume or concentration of their flowers (Hainsworth and Wolf 1976, 43 Nicolson and Fleming 2003, Gutierrez et al. 2004, Zambon et al. 2020), and often for 44 amino acids (Hainsworth and Wolf 1976). Although, the best sources for amino acids in 45 hummingbirds are insects (Abrahamczyk and Kessler 2015). 46 The activity of flower visitors can be predicted by flower phenology (Feinsinger 47 1980, Stiles 1980, Murcia 1996, Rotenberry 1990, Gutierrez and Rojas 2001, Dante et 48 al. 2013, Magilanesi et al. 2014, Gonzalez and Loiselle 2016). For example, movements 49 of hummingbirds are known to be associated with flower blooms (Schuchmann 1999). 50 In temperate forests, hummingbird diversity correlates with flower density, such as in 51 Mexico (Martinez del Rio and Eguiarte 1987), Canada (Inouye et al. 1991), and the U.S. 52 (McKinney et al. 2012). Furthermore, seasonality in the tropics is highly influential in 53 plants and their pollinators (Cruden et al. 1983); temperature and precipitation influence 54 local bird activity (Bourgault et al. 2010) such as foraging time and visitation rates of 55 hummingbirds (Fonturbel et al. 2015). 56 In different tropical forests, several studies have shown an association of 57 nectarivorous birds with nectar resources. Some examples of hummingbirds and their 58 preferences by region are as follows: In Costa Rica - breeding, molt, diversity, density, 59 and movements with blooming of their flowers (Stiles 1978, 1985, Wolf et al. 1976); in 60 Puerto Rico - visits to flowers depend on bill size and corolla length (Kodric-Brown et al. 61 1984); in Bolivia - richness with flower availability (Abrahamczyk et al. 2011); and in 62 Colombia - life cycle with nectar energy and seasonal abundance of flowers (Gutierrez 63 et al. 2004, Cotton 2007, Toloza-Moreno et al. 2014). However studies that looked to 64 find a relationship between hummingbirds and nectar in a landscape gave different 65 results (Ortiz-Pulido and Rodriguez 2011). Other nectarivorous birds may select flowers 66 by traits other than nectar such as accessibility or inflorescence size; that is the case of 67 African sunbirds (Schmid et al. 2015). For hummingbirds, the different foraging 68 strategies (territorial or traplining) are also important in their floral selectivity (Feinsinger 69 1976). 70 The study of nectarivorous bird communities in the neotropics provide 71 opportunities to understand ecological interactions in different ecosystems (e.g. 72 Rodriguez-Flores et al. 2012, Maglianesi et al. 2014) and test specific hypotheses on 2 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.22.262964; this version posted August 24, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 3 73 the drivers of these interactions, such as morphological mismatch (Vinzentin-Bugoni et 74 al. 2014) or nectar quality and quantity (Maruyama et al. 2014). An understudied 75 ecosystem that has an abundant nectarivorous bird community occurs in the upper 76 montane forest of the Andes (Ramirez et al. 2007, Gonzalez 2008). In these forests, a 77 diverse suite of hummingbirds and flowerpiercers is abundant (Gonzalez et al. 2019). 78 However, which factors explain the patterns of plant visitation is little known in this 79 system. Consequently, in this study the question is: Which traits of flowering plants are 80 associated with visits by common nectarivorous birds? I hypothesize that traits 81 associated with energy explain flower visits better than other floral traits. 82 83 Methods 84 Study Area 85 This research was conducted in the elfin forest in Unchog, located in the high 86 Andes of Peru (9° 42' 32.33" S, 76° 9' 39.13" W; 3700 m) from 2011 to 2014. The elfin 87 forest is considered as an ecotone between the cloud forest and the puna grassland. It 88 has a marked seasonality of dry (May to September) and wet periods (October to 89 March). The dry season is not devoid of rain, but it has less rain than the wet season. 90 The temperature is cold, colder in the dry season, and the annual range varies from -1 91 to 15oC. 92 The landscape in Unchog is hilly, with small forest pockets dominated by 93 Weinmannia. The non-forested area is a matrix of puna grasslands with shrubs, the 94 most common one being Brachyotum spp. I sampled three sites that concentrated the 95 most extensive groves of elfin forest (~8 ha each), embedded in an area of 300 ha. 96 These sites ranged from 0.6 to 1.7 Km from each other. The plant composition was very 97 similar in the three sites (Sorensen index of similarity ranged from 0.72-0.80 among 98 sites), so I pooled all the information on plant traits. 99 100 Study Species 101 Nectarivorous birds present in the area were recorded by direct observations with 102 binoculars. I walked inside the forest patches and forest edges, recording the birds and 103 their visits to the flowers. I considered a visit as the moment when a bird fed on a flower 104 or flowers of a plant, disregarding the number of flowers visited and if the visit was 105 legitimate (pollinating) or not because this research considers the visitor’s perspective. 106 A matrix of observed interactions, accounting for the times a bird was visiting a plant 107 was constructed (Gonzalez and Loiselle 2016, Gonzalez et al. 2019). Birds and plants 108 of the bird-flowering plant visitation network that were more abundant and more 109 connected were selected to examine which plant traits predict bird interactions (Ortiz- 110 Pulido and Vargas-Licona 2008). 111 The three most quantitatively important bird species that visited flowers were 112 Coppery Metaltail (Metallura theresiae) – hereafter, the Metaltail; Great Sapphirewing 113 (Pterophanes cyanopterus) – hereafter, the Sapphirewing, and Moustached 114 Flowerpiercer (Diglossa mystacalis) – hereafter, the Flowerpiercer. The Metaltail is a 115 small-billed, territorial hummingbird that weighs 5.07±0.09 g. and has a bill length of 116 12.03±0.87 mm. The Sapphirewing is a large and non-territorial hummingbird with a 3 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.22.262964; this version posted August 24, 2020.